[R-meta] question regarding metafor and OR
Sandro Marini
sandro.marini08 at gmail.com
Thu Sep 28 16:58:06 CEST 2017
Thank you for the clarification.
Best regards,
Sandro
On Sep 27, 2017 17:16, "Viechtbauer Wolfgang (SP)" <
wolfgang.viechtbauer at maastrichtuniversity.nl> wrote:
Dear Sandro,
I assume you mean that you have fitted a model with *log odds ratios* and
now you want to fit a model directly to the odds ratios.
Why would you want to fit a model to the odds ratios? Odds ratios are not
symmetric around 1, so results are going to be non-sensical (the average of
OR = 2.0 and OR = 0.5 is not 1 but 1.25, so even though the two odds ratios
are exact opposites of each other, their average yields a value above 1).
Also, odds ratios do not have a normal sampling distribution (not even
approximately), so model assumptions are therefore violated.
Best,
Wolfgang
-----Original Message-----
From: R-sig-meta-analysis [mailto:r-sig-meta-analysis-bounces at r-project.org]
On Behalf Of Sandro Marini
Sent: Wednesday, 27 September, 2017 17:54
To: r-sig-meta-analysis at r-project.org
Subject: [R-meta] question regarding metafor and OR
I have a model with Odds Ratio and their 95%, I have also the estimates and
their standard errors
So far I have used the estimates = *yi *and the standar errors = *sei *in
the *rma* function.
But what if I wanted to use the Odds Ratio? how I should define the *vi* or
the *sei* arguments in the same *rma* function?
Thank you for your help,
Sandro
[[alternative HTML version deleted]]
More information about the R-sig-meta-analysis
mailing list